If you want to run this directly using Python or develop the tool further, first execute the following commands to get started with pipenv. Make sure you're using at least Python 3.9

pip install --user pipenv
pipenv install

then you should initialise the inventory and ssh-data submodules where all the FSFE-specific data is stored.

git submodule update --init --remote inventory
git submodule update --init --remote ssh-data

finally run to start the UI:

pipenv run python admin.py

and open http://localhost:5000/admin. There you'll find instructions on how to proceed. It boils down to the following:

Configure keys and the systems to which they have access. This project uses the following concept to ease configuration:

Systems are user-host-combinations, e.g. exampleuser@exampleserver.com. For each system the location of its authfile needs to be specified, e.g. /home/exampleuser/.ssh/authfile

Profiles are simply a utility. They allow the user to specify which keys have access to which system in a more straightforward way than adding every key to every system individually. One can for example specify an admin group which has access to all root accounts on all hosts. Then adding a new key that is supposed to have this level of access can just be added to the admin profile.

Keys are simply public SSH keys. They consist of a keytype (e.g. ed25519), the keystring, a comment and an option.
Once everything is configured correctly, you can generate the variables that the Ansible playbook at the root of this project needs in order to carry out its function, i.e. updating all the ~/.ssh/authorized_keys files on all hosts. Generate the needed Ansible configuration files by clicking the button labelled Generate Ansible Config.
At this point, All that is left to do is to deploy all the keys via a slim ansible playbook using the command below:
```
ansible-playbook playbook.yml
```

Architecture

UI

A simple CRUD interface based on Flask Admin based on data model managed in SQLAlchemy (a powerful ORM implementation)

Database

A simple data model with three tables (key, profile, system) and two association tables (at_key_profile, at_profile_system).

Deployment

A pre-deployment python script which creates and the correct variables for each host and the contents of the keystringfiles directory. This script is called from the homepage of the UI, i.e. /admin by clicking the aforementioned button.